scorecardresearch Skip to main content
Ideas | Linda Rodriguez McRobbie

When babies felt no pain

Alex Kingsbury, Heather Hopp-Bruce, Dante Ramos/Globe staff; apimages

Up until the mid-1980s, infants didn’t feel pain.

For years, even as life-saving surgeries became more invasive, longer, and more intense, the majority of newborns still underwent them without anesthetic. Often, they were given nothing more than a muscle relaxant to keep them from thrashing around during the operation. It was the best practice at the time, following the best available science. And it was horrifyingly wrong.

Pain — emotional, spiritual, physical — is fundamental to the human experience. Yet we are still struggling mightily to understand it. Pain is unique to the individual, but also, crucially, in the eye of the beholder. It can defy language’s ability to convey and science to fully explain, and it’s a problem that appears and reappears in issues from the opioid crisis to abortion.


It probably goes without saying that infants can feel pain, as any parent or pediatrician could tell you. But it wasn’t until 1987 that the American Academy of Pediatrics formally declared it unethical to operate on newborns without anesthetics. Why physicians would knowingly inflict pain on newborns is complicated, but there were legitimate concerns that anesthesia itself could harm or kill the child.

More significantly, however, the medical establishment had convinced itself that babies couldn’t feel pain. Because babies can’t speak for themselves, scientists were left to interpret their body language. Studies from the 1940s supposedly confirmed that infants hadn’t yet developed the neurological capability for pain because they didn’t seem to react to pinprick tests.

Cloistered in medical journals, subsequent studies on infant pain showed that their pain responses are as well-developed as older children and that even fetuses, from the third trimester of gestation, possess the systems necessary to feel pain. But few surgeons and anesthetists knew about them, and challenges to accepted wisdom went ignored.

Change was — is — slow to come. A 2003 study found that though most major surgical procedures were now accompanied by analgesic relief, only around a third of newborns were given analgesics for routine yet painful procedures, including blood tests and circumcision. And though we know now newborns probably experience pain “more intensely than older children or adults” and that exposure to pain and physiological stress can alter central nervous system development and pain sensitivity through adolescence, we’re still not getting it right. A 2013 survey of nurses by the National Association of Neonatal Nurses found that fewer than half of them felt that neonatal pain was well managed.


At the heart of this failing is a basic misunderstanding: It’s still pain, even if it doesn’t look like it. In fact, it’s not just infants whose pain is invisible to doctors, who struggle to see it in the population at large. If only there were some objective way to measure it.

Study after study shows that humans are not very good at recognizing that the pain that other people report is real. Physicians have good reasons to suspect their patients’ self-reports of pain, especially in the midst of an opioid crisis fueled by rampant over-prescription of pain-killers. But more often, humans misunderstand others’ pain because of inaccurate preconceived ideas about what pain responses look like.

Not only is the diagnosis of pain complicated by bad science — as in the case of newborns — it is also distorted by our own biases. In 2001, a paper “The Girl Who Cried Pain: A Bias Against Women in the Treatment of Pain” in the Journal of Law, Medicine, and Ethics, found that women are more likely to seek treatment for chronic pain, but are also more likely to have their verbal reports disregarded by clinicians. According to a 2016 study, black children suffering from appendicitis are less likely than other races to receive pain medication, including for severe pain.


Compounding the part our biases play is that much — too much — relies on an inaccurate method of measuring a subjective experience. The gold standard of pain assessment remains self-report, a pain rating on a scale of one to 10. Reporting the “right” number can mean the difference between an insurance payout or not, physical therapy or not, medication or not. But that’s a maddeningly imprecise way to do medicine, and it’s no good if people can’t communicate honestly (those addicted to painkillers) or at all (as with those suffering from locked-in syndrome, dementia, stroke, or mental illness, or any of the other myriad other conditions).

The inability to recognize pain when we see it has enormous implications for some of our most contentious debates. Abortion opponents, for instance, have pushed laws in numerous states claiming that fetuses can feel pain at 20 weeks, justifying a ban on abortion after that point. (The American Medical Association, basing its argument on the neural and synaptic development of the fetus, contends that fetuses don’t feel pain before the third trimester, 28 weeks.) And as the uncertainty around the future of healthcare in America rolls on, as the opioid crisis seems to deepen, as the number of Americans reporting severe chronic pain tops 50 million, it’s clear: We need a better way to see pain.

The best candidate so far? Figuring out where pain lives in the brain. It’s also proving the most contentious.

The increasing sensitivity and availability of magnetic resonance imaging and functional MRI, which can be used to measure blood flow and therefore activation in the brain, have led to an explosion of studies purporting to reveal the neurological circuitry of everything — love and sexual desire, jealousy and schadenfreude, hate. The idea of locating the wellspring of human behavior and emotions in the dense wiring of our brains has a neatness that we naturally appreciate. But an objective biological signature for pain could help improve health outcomes for sufferers, from the simple acknowledgment that their pain is real to targeted drug treatment therapies to alleviate it.


The pain matrix is a group of specific areas of the brain that respond to painful stimuli. Findings have consistently included the anterior cingulate cortex, somatosensory cortex, thalamus, and portions of the central nervous system in that matrix. The pattern of how these areas are activated is considered a potential candidate for a biomarker of pain. That is, if you believe it really is only pain the network is responding to — and not every pain researcher does.

Tor Wager is the University of Colorado neuroscientist whose 2013 paper in the New England Journal of Medicine claiming to have found pain’s “neurologic signature” — a specific and predictable pattern of brain activity across multiple regions correlated with pain — sent up big sparks in the pain research community. In the original study, he said, he and his colleagues were able to determine a pattern of brain-wide activity that reliably tracked with increasing heat stimulation.


Since the initial study, Wager says they have replicated their findings in other settings, as well as teased out the pattern in which it had been entangled with other emotional experiences. In 2015, a student in Wager’s lab was able to show how physical pain and the pain of social rejection, previously thought to be so similar as to be the same, follow two different neural circuits. “I’m very confident that we have a measure that responds to many types of evoked pain, heat shock, mechanical, laser, pressure, cold, visceral stimulation, rectal, esophageal, gastric, it responds to evoked painful stimulation in all of these,” he said. “It has a sensitivity and specificity profile across multiple types [of stimulation] that’s better than I would have imagined when I started to do this work.”

Pain is a subjective experience, but that doesn’t mean that there aren’t observable, objective processes driving that experience. “I think that we’re observing the processes that go into creating that experience,” Wager explained, cautioning, “but we can’t measure pain directly. . . What it is intended to be is a measure of a component of pain. You’re not measuring pain, you’re measuring a physiological system that people often correlate with pain.”

Other scientists have found fault with Wager’s research. Tim Salomons, a pain researcher and professor of psychology and neuroscience at the University of Reading in Britain, is co-author of a study that questions the integrity of the pain matrix model. Published in 2016, his experiment looked at the pain matrix responses of two people born without the ability to feel pain and four control subjects. The volunteers underwent “noxious mechanical stimulation” (a sharp needle pressed to the skin, but not breaking it — not unlike the 1940s pinprick tests) while undergoing an fMRI scan. Salomons and his colleagues found that the brain activity responses of all subjects were largely indistinguishable, meaning that the pain matrix was activated in people in whom a pain response was impossible.

“What we generally see when we do experiments, this so-called pain matrix lighting up — in order for that to be the measure of pain, you have to make the assumption that it really is pain that we’re measuring,” he explained. Based on his research, however, “it really can be any salient sensation” that activates the areas of the pain matrix.

Stephen McMahon, one of Europe’s leading pain researchers and a professor of physiology at King’s College London, agreed. “Something new is happening, not painful, but new. What we can’t say is that the activation is a signature for pain,” he noted.

Moreover, he said, there is a fundamental flaw in using fMRI to understand how pain works. “The fMRI, trying to look into people’s brains, is a brilliant technique. But it’s not very good at determining causality,” he explained, adding that brain imaging science tends to suffer from “reverse inference” — seeing patterns of activity in the brain and linking that to a cognitive process, rather than the other way around. “When two things are correlated, you can’t conclude that one causes the other.”

Wager dismissed Salomons’s study as “kind of a straw man,” saying that the experiment looked only for activation of the areas implicated in the pain matrix, not the specific signal his team found, nor did it examine the magnitude of activation. The sheer amount of data he’s producing in support of a pain signal might also help resolve the problem of correlation not being causation. But the divide between those who believe that pain has a neurological signal and those who don’t is more fundamental.

Both camps agree that pain is a complex cognitive phenomenon. Pain is a multilayered experience, dictated by not only the painful stimulus itself — the cut thumb, the broken leg, the crowning baby — but also context, the individual’s mental state, needs, wants, expectations, personal physiology, group dynamics. Studies have shown that the intensity of pain can be mitigated by laughing, or by holding someone else’s hand. Other studies have shown that when people expect to feel pain, say after a surgical procedure, they generally do. People can even feel pain that doesn’t belong to them — in studies that employ the “rubber hand illusion,” for example, people come to feel ownership of a mannequin hand after a series of visual and tactile stimulations. Researchers are able to induce and reduce sensations of pain in the subject’s real hand by burning the fake hand and later applying a topical analgesic, again to the fake hand.

But for Wager, undergirding the pain experience is a predictable pattern of activation in the brain that occurs when pain is happening. What the brain then decides to do once this activation has taken place can dictate whether that sensation is read by the individual as pain. “I can do things that will change the pain that you report to me but not affect the pain signature,” Wager said. This means that a person could report not being in pain, but the pattern of activation in their brain would say otherwise.

This makes Salomons uncomfortable. Can you still call it pain if it doesn’t hurt? “When we are looking for an objective measure of pain, we’re kind of presuming that there is some observable, discrete biological process that’s pretty much interchangeable with pain. If this process occurs, it is both necessary and sufficient for the experience of pain, and that’s where things start getting difficult,” he said. “Pain is always going to be some kind of interaction between the body and the brain — I don’t know if it will be reducible to a set of neural processes.”

Wager doesn’t see his work as reductionist, noting, “Our emotional lives are so much more complex and nuanced that what we can track in the brain, but that doesn’t mean that we can’t find things in the brain that are important and reliable to understand emotions.” And some of those who support Wager and the pain matrix don’t see any incongruity between the existence of a pain signature and the complexity of how we experience pain. “When I squeeze just about anybody’s finger, those nociceptors will be activated, some early perceptual areas are going to be activated in the brain that say there is a nociceptor firing. But what you do with it after that is subject to a lot of things,” explained Greg Siegle, a neuroscientist at the University of Pittsburgh who uses neuroimaging to understand emotional cognition in people with mood disorders.

Siegle offered a comparison: There is a part of the brain that processes the smiles of other people. Say you see someone on the street smiling. How you feel about that smile is dictated by what you know — if, for example, you knew that the person smiling had just saved a drowning kitten, you might genuinely smile back. If you knew they had just drowned a kitten, you might not. “There’s a smile network that activates both times, but your experience of that smile will be entirely different based on the context,” Siegle said. “Pain is like that. There is a pain network, but there is all sorts of psychological mediation that goes on once that pain network is activated.”

Practical applications of the pain signature are still some ways off. Problematically, however, some people do want to act on the science now. For example, an American company, Millennium Magnetic Technologies, is selling its services as a kind of lie detector for pain, a claim that critics say is dangerously premature. The company has already been called on in several pain-related lawsuits to use an fMRI to determine whether a pain signal is present in people suing for damages. McMahon was recently asked by a British high court whether this kind of scan should be admissible evidence. Absolutely not, he responded. “That is just pie-in-the-sky at the moment,” he said.

Wager agrees that there is “a lot of reason to be really cautious,” echoing some of Salomons’ concerns that using brain imaging as part of diagnostics might result in patients’ pain being less believed, not more. “But the question is, where do you go from there? Do you stop doing neuroscience of pain? Or do you come up with what this means and doesn’t mean. . . what are the limits of it, what does it work for?” he continued. “Any brain measure or biological measure, any technology can be misused. . . what we have to do is be smarter and more thoughtful and compassionate about how we use them. . . . We do have the potential to make people’s lives better, we just have to be smart about it.”

As worrying as the potential for getting it wrong is, equally concerning is the idea that we might be missing someone’s pain simply because we can’t see it — just ask the millions of parents who watched their babies go under the scalpel without anesthesia.

Linda Rodriguez McRobbie is an American freelance writer living in London.